Bioscience Method 2024, Vol.15, No.1, 21-27 http://bioscipublisher.com/index.php/bm 21 Scientific Review Open Access Unveiling the Mechanism of Proprioception in Primates: The Application of Task-Driven Neural Network Models Natasha Liu Cuixi Academy of Biotechnology, Zhuji, 311800, China Corresponding email: natashaccliu2023@gmail.com Bioscience Method, 2024, Vol.15, No.1 doi: 10.5376/bm.2024.15.0003 Received: 01 Jan., 2024 Accepted: 03 Feb., 2024 Published: 14 Feb., 2024 Copyright © 2024 Liu, This is an open access article published under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Preferred citation for this article: Liu N., 2024, Unveiling the mechanism of proprioception in primates: the application of task-driven neural network models, Bioscience Method, 15(1): 21-27 (doi: 10.5376/bm.2024.15.0003) The paper titled "Task-driven neural network models predict neural dynamics of proprioception" was published in the journal Cell on March 21, 2024, by authors Alessandro Marin Vargas, Axel Bisi, Alberto S. Chiappa, Chris Versteeg, Lee E. Miller, and Alexander Mathis, are from the 1Brain Mind Institute, School of Life Sciences, Ecole Polytechnique Fe´ de´rale de Lausanne (EPFL), 1015 Lausanne, Switzerland; Department of Neuroscience, Feinberg School of Medicine, Northwestern University, Chicago, IL 60611, USA, and others. This study delved into the neural dynamics of proprioception in primates under active and passive movement conditions through the establishment of task-driven neural network models. Utilizing synthetic muscle spindle inputs and musculoskeletal modeling techniques, the research team simulated the proprioceptive process in animals and trained neural networks to solve multiple computational tasks, testing various hypotheses regarding proprioceptive processing. These models were used to predict the neural activity in the cuneate nucleus (CN) and the primary somatosensory cortex (S1) of non-human primates, thereby assessing the effectiveness of various hypotheses in explaining these neural dynamics. 1 Experimental Data Analysis The experiment utilized OpenSim musculoskeletal modeling and deep learning technologies to generate synthetic proprioceptive inputs, simulating the movement process in animals. Key results include the performance comparison of different neural network models, such as TCNs and LSTMs, in predicting neural activity, as well as the performance differences of these models under active and passive movement conditions. The Results showed (Figure 1) that a normative framework for studying the neural coding of proprioception. By utilizing synthetic muscle spindle inputs and musculoskeletal modeling, neural networks were optimized to solve 16 computational tasks, testing various hypotheses based on learned representations. The study evaluated which type of hypothesis could better explain the neural activity in the cuneate nucleus (CN) and primary somatosensory cortex (S1, area 2) of non-human primates during active and passive center-out reaching tasks. The model types compared include a baseline linear encoding model that directly predicts neural data from experimental data, and three types of neural network models: data-driven models trained end-to-end on experimental data, untrained neural network initializations tested directly on experimental data, and task-driven models trained on synthetic data to perform computational tasks. It can be seen that (Figure 2) by enhancing 2D character trajectories and projecting them into 3D space, synthetic proprioceptive inputs are simulated using a two-link four-degree-of-freedom (DoF) arm model. Muscle lengths and velocities are calculated from the 3D trajectories using inverse kinematics and musculoskeletal modeling. The figure shows the distribution of joint angles in the behavioral data (above) and synthetic data (below), with the motion statistics of the synthetic dataset designed to encompass the biological movements of non-human primates during center-out reaching. Sixteen computational tasks were designed, reflecting hypotheses about proprioceptive processing, each containing one or several learning objectives. Different neural network architectures were
RkJQdWJsaXNoZXIy MjQ4ODYzMg==